Resources

HW

Your own computer (best for debugging)
GPU cluster (recommended for longer training)
Google Colab
- Allows using GPU runtimes for 12 hours a day (also TPUs with Tensorflow)
- Allows to link Google Drive with faculty account

Neural networks Stanford YouTube lectures
Neural Networks - 3Blue1Brown YouTube playlist
Deep Learning - Josh Starmer YouTube playlist
Neural Networks Visualized - vcubingx YouTube playlist
Understanding KL Divergence - Adian Liusie YouTube video
Understanding Cross-Entropy - Adian Liusie YouTube video
A Short Introduction to Entropy, Cross-Entropy and KL-Divergence - Aurélien Géron YouTube video
Building an autograd library - Andrej Karpathy YouTube video